Page 3 of 3

Re: 2 New CTF maps created +1 more.

Posted: Fri Sep 09, 2016 10:57 pm
by Barbie
At least I was able to reproduce that effect on every (?) swjumppad: face to its destination and walk/run in in a way, that your speed is close to 0 when touching its collision cylinder. Then the swjumppad calculates the correct velocity vector to the target, applies that to the Pawn, the Pawn has the assigned velocity (verified by debugging output), but the Y component of that velocity seems to be ignored from the engine that is doing the movement finally: the pawn (me^^) moves along X and Z axis only, but not along Y axis.

Even if I skip the part of calculation the velocity vector that leads to the target and assign a constant vector to Pawn's velocity, the Y component is sometimes ignored and sometimes not:

Code: Select all

simulated function vector CalcVelocity( Pawn Other ) {
[...]
    // commented the next line out: 
    // Other.Velocity = vel;
    // and added this instead:
    Other.Velocity = vect(-115,-1677, 1034);
[...]
}

simulated event PostTouch( Actor Other ) {
    P = Pawn(Other);
[...]
    CalcVelocity( P );
[...]
}
Are there conditions under that a direct assignment to Pawn.Velocity does not work or work only partially?

Re: 2 New CTF maps created +1 more.

Posted: Sat Sep 10, 2016 1:42 am
by Gustavo6046
Barbie wrote:At least I was able to reproduce that effect on every (?) swjumppad: face to its destination and walk/run in in a way, that your speed is close to 0 when touching its collision cylinder. Then the swjumppad calculates the correct velocity vector to the target, applies that to the Pawn, the Pawn has the assigned velocity (verified by debugging output), but the Y component of that velocity seems to be ignored from the engine that is doing the movement finally: the pawn (me^^) moves along X and Z axis only, but not along Y axis.

Even if I skip the part of calculation the velocity vector that leads to the target and assign a constant vector to Pawn's velocity, the Y component is sometimes ignored and sometimes not:

Code: Select all

simulated function vector CalcVelocity( Pawn Other ) {
[...]
    // commented the next line out: 
    // Other.Velocity = vel;
    // and added this instead:
    Other.Velocity = vect(-115,-1677, 1034);
[...]
}

simulated event PostTouch( Actor Other ) {
    P = Pawn(Other);
[...]
    CalcVelocity( P );
[...]
}
Are there conditions under that a direct assignment to Pawn.Velocity does not work or work only partially?
Maybe setting the Acceleration to 0 or maybe setting the velocity like "Other.Velocity *= -1" should work.

Re: 2 New CTF maps created +1 more.

Posted: Sat Sep 10, 2016 3:19 pm
by sektor2111
More confusions come as long as that swJumpad was the best stuff for A.I. and Human jumping as long as you don't mess setup with KICKER around. Also you are able to control a bit your "landing". I was playing maps from 300K (open your eyes there) where swJumpad is used (Wootious map included) and if you play normally (RESPECTING 4 TOUCHERS rule by Unreal Engine) stuff is pretty better than any sort of mess supposed to do A.I. jumping but doing nothing. Aside feel free to combine LE-LC-LE if you can manage that. I prefer to leave alone good stuff because a single fail from 200 or whatever might be caused by a SHOOT changing Velocity or whatever factor (nearby KICKER might screw up) because Kicker has nothing to do with navigation, but is bugging A.I. properly. I was falling down in map which I privately sent at Terraniux because those clowns XAN_A and XAN_B were against me in the same team hitting me in air and bugging my velocity ON Y AXIS :loool: . When they somehow were busy to respawn behind the rocks I felt more comfortable at any landing. Else sWJumpad which I'm using in the middle is doing the job properly even if see a LOG warning (not ScriptWarning) because angle is too closer to 90 degrees, but I still can reach in upper location ( And BOTS too ). In the Level pathed correctly I see a good difficulty for keeping you warmed up else I have a few risks to sleep on keyboard.
As for pressing <walk> while flying, I did not test that, but if things goes nasty the same things might occurs at kicker too...
Edit: A little test Looks like walk and back pressed shortly might screw Y velocity and you might fall at middle of road. Who need WALK Back in that moment ??

Re: 2 New CTF maps created +1 more.

Posted: Sat Sep 10, 2016 3:49 pm
by papercoffee
sektor2111 wrote:Who need WALK Back in that moment ??
Who ever likes to fall down midair :ironic:

Re: 2 New CTF maps created +1 more.

Posted: Sat Sep 10, 2016 4:10 pm
by OjitroC
I've played this map quite a bit now, as well as testing it without bots, and would say that the problem with the swjump pads occurs around 1 in 10 times they are used. IMHO it is particularly noticeable with the jump pad from the central island to the blue base. Hitting the jump pad slightly off centre is more likely to result in a correct jump than hitting it dead centre, I think. Pressing or not pressing the forward key after hitting the jump pad does not seem to make a difference though I haven't tested that much.

The problem is not enough to put me off playing the map though.

Re: 2 New CTF maps created +1 more.

Posted: Sat Sep 10, 2016 8:27 pm
by EvilGrins
Lately I've been in a very what-ctf-can-i-put-utdmt-on mood. As to yours in this thead:
· CTF-DONUT: Not enough room for them.
· CTF-BackAlly: Probably not a good idea, too cramped.
· CTF-CrystalRocks: Could work but I doubt it.

Re: 2 New CTF maps created +1 more.

Posted: Tue Sep 13, 2016 3:40 am
by Barbie
I guess I've found the reason for that vanishing velocity in Y direction: the JumpPad is centred on a square plateau (a 256x256x16 additive brush half in the ground) that is 8 UU above the ground. The collision radius of the JumpPad is 128 so that there are four points where the collision cylinder of the JumpPad reaches the outer edge of the plateau. If you come close and you are still on the ground and not on the plateau, you have a small step in front of you. If the JumpPad is touched in that moment, the velocity is applied to the player, but the step "eats up" the orthogonal component of the velocity.
Possible solutions:
  • reduce collision radius of JumpPad or enlarge plateau (or both ;) until JumpPad.CollisionRadius + TournamentPlayer.CollisionRadius < minimum(SizeX(plateau), SizeY(plateau))
  • lower the step hight or use brush sinking
  • use a sheet instead of a brush for the plateau
  • ...

Re: 2 New CTF maps created +1 more.

Posted: Wed Sep 14, 2016 3:14 pm
by sektor2111
So I was right about swJumpad as it doesn't include issues, issues being excluded. Bugs are coming from SETUP not the code.

Re: 2 New CTF maps created +1 more.

Posted: Wed Sep 14, 2016 3:31 pm
by Terraniux
. This is the first time I use SWJumppad so... Thank you very much Barbie :gj:

Re: 2 New CTF maps created +3 more.(new)

Posted: Mon Oct 03, 2016 8:01 pm
by Terraniux
Added 2 more maps, Warboss and SOTQ.

SOTQ is for modded servers, Warboss has all weapons.

enjoy

See first post for details.